Playing at the edges: use of playground spaces in South Australian primary schools with new arrivals programmes

Resettlement in Australia can be a stressful experience for migrants who are largely expected to quickly assimilate into the broader Australian culture. This expectation is heightened for those people who arrive as refugees, and particularly those who arrive without humanitarian visas and must endure mandatory detention. For both migrants and refugees, attempts at establishing themselves or developing a sense of belonging to their new community are likely significantly hindered by the welcome (or otherwise) they are provided with and the terms upon which inclusion is o ffered. This paper uses a multi-method approach to examine how these issues of belonging and inclusion played out in the use of space in two South Australian primary schools that include a programme (NAP) for new arrived students. The paper considers the way in which NAP and non-NAP students utilise playground spaces, and compares and contrasts these observations with the views of teachers at the schools. Specifically, the findings indicate that NAP students were largely relegated to the margins of the playground and experienced di fficulty in claiming school spaces as their own. The paper concludes by making suggestions for schools on the basis of the findings, with a focus upon examining the power relations that exist between NAP and non-NAP students and the role of schools in developing a global understanding of inclusion and exclusion.


Introduction
The arrival of refugees without visas in Australia has long been a contentious public issue, which has been exploited and misrepresented by both politicians and the media.Much of the rhetoric surrounding refugees who arrive in this way concerns issues of "integration" or "absorption", together with a focus on the strain on resources and space that such people might supposedly cause (notwithstanding the fact that the number of refugees arriving "illegally" without visas in Australia remains small compared to total immigration levels in Australia, and to other people in Australia "illegally" such as people who over stay their visas).This rhetoric, combined with fear campaigns promulgated within the political Correspondence to: C. Due (clemence.due@adelaide.edu.au) and media arenas, has led to mandatory detention policies, and the governance of space to excise certain areas of Australian territory so that refugees arriving by boat on islands off the Australian coast are unable to claim asylum in Australia.
This control of Australian space has created a situation in which resettlement for refugees, both legal and "illegal", can be a difficult process, in which they may have little choice about their site of placement and little knowledge of the services and assistance available (or unavailable) to them, leading to potential difficulties in securing employment, establishing a household and settling into the community into which they are placed (Mares, 2001).Yet despite these challenges, the primary message made available to refugees is the need for integration into the local community and the adoption of "Australian values" (see, for example, the booklet "Becoming an Australian Citizen (2007) published by the Australian Government for newly arrived immigrants and Published by Copernicus Publications.
C. Due and D W. Riggs: Use of playground spaces in South Australian primary schools with NAP refugees).Furthermore, the onus is generally placed upon refugees to undertake the bridge building work of engaging with non-refugees, rather than the community into which refugees are placed accepting some responsibility for developing strategies to bridge divisions which may exist between refugee and non-refugee communities (Morrice, 2007).
Research regarding the experience of refugees in Australia continues to find that newly arrived people and families often struggle to adapt to their new home (especially refugees from African countries who are highly "visible" in the communities into which they are resettled, see Farhadi and Robinson, 2008).In many instances this is due to discrimination from within the broader community, which is fed by media and political rhetoric in which refugee communities are depicted as "ghettos" that are intentionally isolated from mainstream communities (for examples of analysis of this rhetoric, see O' Doherty and LeCouteur, 2007;Perera, 2007;Sidhu and Christie, 2002).Indeed, such rhetoric was seen in comments made in 2007 by then Immigration Minister Kevin Andrews regarding the supposed failure of African refugees to integrate into the broader Australian community, and the subsequent media response to this (Due, 2007).
This brief outline of the current situation of refugees in Australia illustrates well the fact that the use of space, and claims to place, are never neutral: the spaces in which all people move are highly regulated by social norms that determine how individuals can move, and which shape the types of engagement that are possible (Holloway and Valentine, 2000a).Here, it is useful to consider Sack's concept of territoriality, which provides a framework through which to understand the regulation of space.Sack argues that "territoriality in humans is best understood as a spatial strategy to affect, influence or control resources and people by controlling area" (Sack, 1986:1), and that this control of area primarily revolves around power (or the lack thereof).Specifically, and in relation to the present paper, it is important to recognise the fact that such control often rests upon the racialisation of space in ways that centre the values of dominant groups, and which reinforce binaries of inside and outside in order to legitimate exclusionary practices (Trudeau, 2006).How refugees move in Australia, for example, is highly managed by governmental agencies at a broad level, and micromanaged by non-refugee (predominantly white) Australian citizens who feel entitled to act as managers of the spaces in which they move, often to the detriment of refugees and other migrants (Hage, 1998).
In a similar way, the use of space and the norms of behaviour that operate within that space are highly regulated for children.Previous research in the area of childhood geographies and spatiality has investigated the ways in which children's spaces are used by children themselves, and are managed and supervised by adults who are keen to control, protect and socialise children (Holloway and Valentine, 2000b).This is relevant to the home, public and school environments.For example, Smith and Barker discuss the ways in which power struggles over spaces play out between boys and girls in out of school care, arguing that while spaces in childcare centres are designed by adults, they are constantly being re-worked and challenged by the children using them (2000).These findings are reflected in research undertaken by Thomson who, applying Sack's notion of "territorialisation" to school playground spaces, similarly found that children were active in their use of space within the playground, despite strict controls by teachers monitoring the playground spaces during break times (2005).Such agency was seen by Thomson in the way that children would "play" with the boundaries of the school yard by, for example, deliberately walking around the edge of a space they were not allowed to enter.
Nonetheless, playground spaces remain important sites in which children learn about social hierarchies, and encounter cross-sections of society and new people, in some instances for the first time (Kelly, 1994).This may be especially the case for refugee or migrant children entering school spaces, since attending a school may be their first point of contact with their new community.Whilst dominant group children may have some degree of power over the spaces they inhabit, it is arguable that racially marginalised children entering such spaces may not have the same power to resist the norms already in place, or to shape spaces for themselves.Indeed, this was found in research by Thomas (2009) on territoriality and race in LA high schools.Thomas found that the use of space within the schools was highly moderated by social norms that regulated which bodies could move in which spaces.In fact, Thomas suggests that crossing borders within the schoolyard is typically constructed as taboo across all cultural groups and youth sub-cultures, and that it may be especially regulated in relation to crossing racial categories in social contexts where such categories are rendered salient.
Whilst not all schools may be strictly divided down racial lines, a number of studies have found that children as young as three or four understand the concept of race and the power relations inherent within the construction of different racial categories, and in many cases show ability to use racial categories in harmful or discriminatory ways (see, for example, Van Ausdale and Feagin, 2001).Other studies have found that school geographies may be divided down lines of class as well as race, such as Kuriloff and Reichert's (2003) study of an elite boys' school in America, which found that the school was divided along lines of the boys considered "lifers" (boys from upper and upper-middle class backgrounds), and boys from lower socio-economic or non-white backgrounds who were considered to be at the "margins" of the school.Kuriloff and Reichert argue that their study has relevance for public as well as private schools, and clearly both race and class issues would impact upon the ways in which refugee children are able to "fit into" school environments.
In this paper we outline findings from a project undertaken with two South Australian primary schools that include a New Arrivals Programme.New Arrivals Programmes (NAPs) are programmes into which students are placed if they meet the criteria of having low levels of English language skills.Thus the programme is open to students from any non-English speaking background (NESB), provided that they have been in Australia for less than 12 months.Students typically remain in the programme for around a year, or until their English skills have reached a required level, at which time they transition into mainstream classes either at the same school or elsewhere.In this paper we look specifically at how the school spaces in question were used by NAP and non-NAP students, and we compare observations made by the first author with the reported perceptions of teachers in relation to the use of space and the degree of space sharing by NAP and non-NAP students.Importantly, we recognise that the two schools are relatively different in terms of the playground spaces they offer to students, and the demographics of their student bodies.As such, we treat them as two individual case studies for the purpose of this paper in order to draw attention to their differences.Nonetheless, we signal here that our findings suggest relative homogeneity in terms of the gap between teachers' and our own observations of the degree of space sharing that occurs.We conclude by examining some of the possible reasons for why NAP and non-NAP students may not play together, and offer suggestions as to ways in which this may be addressed by schools.

Sample
The study took place in two of the 16 South Australian primary schools with NAPs.In order to preserve their anonymity, we refer to them here as Hills Primary School and Plains Primary School.Hills Primary School (HPS) had a total of 222 students at the time the research took place, with 75 NAP students spread across six NAP classes.As such, NAP students accounted for 34% of the student body at the school.Of the NAP students 29 (39%) were refugees.Hills Primary School was rated as a category six school on the Department of Education and Children's Services (DECS) Index of Educational Disadvantage, where a category one school serves students from the most disadvantaged families, and category seven schools serves students from the least disadvantaged families.School categories are calculated on the basis of a combination of data from the Australian Bureau of Statistics (such as median house prices) and DECS data relevant to the school (such parental income levels).
Plains Primary School (PPS), in contrast, is a category 3 school.It has a larger number of students (294), but is situated on a smaller amount of land, with a smaller playground space.The school has around the same number of NAP students (70), spread over five different classes.As such, NAP students comprise around 24% of the student population.Of the NAP students enrolled in the school at the time the research took place, 18 were refugees (25%).
Both schools had a diverse student population, with students from a wide range of cultural backgrounds, meaning that in terms of numbers, white Australian students were not in the numerical majority (though as we demonstrate below, this did not prevent them from holding a dominant position within the school yard).

Data collection
Ethics approval was granted by both DECS and the authors' University.Consultation meetings were then held with both schools identified as key sites for the study by DECS.Both schools expressed enthusiasm at being involved in the project.Information sheets were provided to parents apprising them of the study, and its minimal impact upon students in their everyday schooling.
Over a period of eight weeks, the first author conducted an ethnographic study of students' play within the school yards of HPS and PPS.These observations focused primarily upon interactions between NAP and non-NAP students, but also examined how students used the playground space.The first author attended each school for 2 days a week over the observation period, in which approximately fifteen minutes was spent each lunch time in a NAP classroom while the students were eating their lunch before observing student play during lunch breaks.Over the research period each NAP classroom was visited two or three times.The first author was introduced to the students by either the class teacher or the NAP co-ordinator as someone with an interest in observing students during play time.During these periods the first author engaged in informal conversations with the students as this time was not an ordered classroom time.The aim of these interactions was to facilitate rapport with students, as well as provide the first author with an opportunity to familiarise herself with students so as to be able to recognise them in the yard.This was important as the research relied upon knowing which students were involved in the NAPs and which were not.
Conversations with students typically took many different forms.In order to initiate a conversation, the first author would often ask a simple question such as what the student's name was, how their day was going, or possibly what they liked to do during playtime.As the aim of this time was not to obtain data per se (but rather to build a level of trust and familiarity), these conversations would often focus on topics determined by the students (who in many cases were eager to chat), and focused on topics driven by the students such as what colour their hair colour was, what they did on the weekend, and the number of brothers or sisters they had.However, in many instances students would speak on topics directly related to the study, such as who their friends were, who they played with and why, and what spaces they liked to use.Notes were not taken at the time of these conversations (in order to facilitate a relaxed atmosphere), but instead were www.soc-geogr.net/5/25/2010/Soc.Geogr., 5, 25-37, 2010 C. Due and D W. Riggs: Use of playground spaces in South Australian primary schools with NAP made once the first author had left the classroom and was in the yard.
During the half hour spent in the yard the first author observed in turn each of the main spaces of the yard in order to ascertain which students were primarily using each space.Maps were used to depict where students were, and a charting system was devised in order to "count" students in terms of how many NAP and how many non-NAP students were using each space during the time the researcher was there.
Extensive field notes were also taken in order to record incidents which took place in an area whilst it was being observed.Such incidents would include what students were doing, any conversations held with students or teachers in the yard, and where exactly within each space students were (for example whether they were playing on the equipment, or running around the edge of the area).Interactions between NAP and non-NAP students were considered especially important.The observational notes were important for adding richness and complexity to the quantitative charting system.
It is important to note here, and as Thomson (2005) suggests, that whilst students were told who the first author was and what she was doing, this is not the same as obtaining informed consent.As the first author simply entered the classrooms and initiated conversations (and subsequently conducted observations in the yard), students at the schools were not given an opportunity to decline to participate.Of course, where a child seemed reluctant to speak to the first author during classroom conversations the first author would end the conversation and move on.However, and as Thomson (2005) states, students within school grounds are typically not in a position whereby they are able to refuse the attention of adults.
In addition to the observations and charting data, a questionnaire was completed by teachers.The teachers received an information sheet detailing the study, and were required to sign a consent form.Two separate envelopes were provided to enable the teachers to return the questionnaire in one and the consent form in the other.These questionnaires were designed to obtain teachers' views on school policies regarding the NAP, the school environment itself, and the use of space in the school yard.The questionnaire was a combination of questions with responses on Likert scales together with open ended questions in which teachers were given spaces to write their views.

Analytic approach
Upon completion of the observation period, quantitative data were extracted from the space-use charts completed by the first author.Accompanying excerpts were taken from the field notes in relation to any reference to the use of space by students, with a focus on details that may not be clearly identified in the broad categories used in collating the quantitative data (i.e., the main spaces in which students played in each school).Data from the teacher questionnaires were col-lated in a similar way, with ratings of space-use tallied across teachers within each school, and short answer responses relating to space extracted from the questionnaires.
Tables with percentages or incidences were constructed along with charts to highlight overall usage of school spaces from the quantified observational data.Significance testing was undertaken on this data using log-likelihood ratio tests to determine the degree of association between spaces and their use by NAP, non-NAP or both groups in combination.Log-likelihood ratio tests were used in preference over the more commonly used chi-square tests due to the low numbers within most cells.Tables were also constructed to report the responses that all teachers in combination gave within each school to Likert scale questions on space use, and charts were generated to highlight the overall rating by teachers of the degree to which NAP and non-NAP students play together.Significance testing was not undertaken to compare observational and teacher data as the measures were not comparable (observational data provided information on the use of space by each group of students on specific days, whereas teachers simply rated the degree of shared use of each playground space as they perceived it in retrospect).
Whilst our focus here is primarily upon the quantitative data, we consider it important to present in the following analysis extracts from the ethnographic observations along with the quantitative data itself.Partly this is so as to give the reader a feel for what was observed by the first author in a qualitative sense, and partly it is to acknowledge the mixed-methods nature of our data: whilst some of the quantitative data was collected as numbers, other aspects of it (as described above) were collected via field notes and coded.Including some information from the field notes thus helps to provide an indication of the validity of our coding system, just as the codification of the field notes allows us to report in broad brush strokes the observational data collected (and to make comparisons across the data through non-parametric significance testing).

Plains Primary School
As can be seen in Table 1, whilst there were many instances observed by the first author of NAP and non-NAP students playing in the same area at PPS, very few of these instances involved the two groups actually playing together.Rather, NAP and non-NAP students typically played in distinct groups with little interaction between the two.
A log-likelihood ratio test was performed to examine the association between student play cohort (NAP or non-NAP) and play location to determine if NAP students and non-NAP students were more likely to play in the same as opposed to different areas.The association between these variables was significant, LRR (4, N = 21) = 20.02,p < .01.NAP and non-NAP students were more likely than not to play in different  As can be seen in Fig. 1, very few incidences of NAP students playing on their own (i.e., occupying an area to the exclusion of non-NAP students) were witnessed across all of the six main play areas (only 2% of all incidences of play by either group).
Examination of the field notes suggested that in addition to sometimes playing in the same space as non-NAP students (though not with them), NAP students often played on the edges of the main play areas, and thus were not visible within the charting system (and its focus upon the main play areas within the school).For example, extracts from the field notes state that "A group of four NAP girls are playing around out the front of their classroom eating their lunch. . .there are no other children playing in this area" and "The NAP students whose class I visited this morning are out the front of their classroom -it looks like it could be almost all of them.When I first walk past they are playing chasey and later they are playing at acrobatics on the poles -holding on with either one hand or both and swinging around.There are both boys and girls here".
These findings reiterate Thomas' (2009) research on play ground spaces which found that such spaces were often divided according to students' cultural grouping.So, for example, at PPS small groups of NAP students (typically 2-4 younger students) were often observed sitting and talking on the edge of the oval, or at the edge of playground areas, but rarely venturing into the area itself.Additionally, NAP students at PPS were often witnessed eating their lunch in the area directly outside their NAP classroom, and then staying there to play (as the above extracts illustrate).These students typically demonstrated little movement across the playground throughout the observation period.
It is worth noting here that whilst Blatchford, Baines and Pellegrini ( 2003) note from their research that boys tend to take up the majority of the playground space (leaving girls to play around the edges), differences were observed in this study between the play styles of NAP and non-NAP girls.Thus, for example, a small number of non-NAP girls were observed playing with boys in the centre of the oval, and non-NAP girls were also observed frequently sitting in large groups in the main spaces of the school.Comparatively, NAP girls rarely played in the main spaces at all, instead preferring areas close to their classroom.
Of course it may be suggested that these findings are a product of the fact that there were relatively few play areas available to all students at PPS due to the size of the yard (as can be seen in Fig. 2), and being in the numerical minority meant that NAP students were unable to stake out an area just for themselves.However, we would suggest that a focus on the number of NAP students (and its role in determining ability to claim a territory) only tells part of the story.It may also be the case that, regardless of their numbers, NAP students would struggle to claim areas all for themselves as the regulation of play spaces tended towards the privileging of non-NAP student play styles over those of NAP students.As such, it could be argued that NAP students are rarely able to control the playground space in order to use it in ways determined by them.An anecdote from the field notes helps to illustrate this: One of the NAP teachers tells me of two Anglo-Indian students in the school (a boy in year seven and a girl in grade one) who are very new to Australia.The little girl has been crying all morning and refused to let her mother drop her off in her classroom and instead would only be dropped off at the office.I ask if there is any reason for why the little girl's behaviour has changed (they have been at the school since the start of term, and until now the little girl had not exhibited such behaviours).
The teacher responds by telling me that "the little girl had been hanging out with her brother in the yard but they were recently told off by the yard duty teacher and told to play separately as generally within the school it is considered inappropriate for year sevens and year ones to be playing together".Later I see them in the yard sitting together by themselves finishing off their food.
www.soc-geogr.net/5/25/2010/Soc.Geogr., 5, 25-37, 2010 Examples such as this highlight the complex ways in which school spaces are managed not only by students in their play behaviour, but also by teachers (Thomson, 2005).
Here a yard duty teacher can be seen to have managed the ways in which two NAP students played on the basis of school rules established to govern all students, with little attention to the specific needs of NAP students.So, for these siblings (and perhaps especially for the young girl), a possible need to play together is overridden by a school policy that prevents students playing together across classes.Miller (1997) reported similar findings in her research on refugee students' experiences in mainstream classes.One of her participants, a 20 year old boy who was placed in a year 10 class on the basis of his English language skills, was reprimanded for having a girlfriend in his year level.Again, this type of treatment of NAP students fails to recognise that such students will have specific needs that differ from those of non-NAP students (for whom school rules were likely primarily designed).
Another example of NAP students not conforming to the play behaviours expected of students at PPS was the fact that NAP and non-NAP students were rarely seen playing together, despite an expression from both schools that they wished for this type of "integration".Indeed, both NAP and non-NAP students were often observed playing across classes, but rarely playing together.Whilst in casual conversation in the yard teachers seemed to acknowledge this difference in play cohorts, data from the questionnaires suggested that teachers perceived much higher levels of shared play between NAP and non-NAP students than was observed by the first author, as can be seen in Fig. 3.
Here we can see that the majority of teachers (62%) at PPS percieved NAP and non-NAP students playing together across all main play areas at least sometimes, with 27% believing this happened frequently.Looking at the main play areas individually, we can further see the difference between the perceptions of teachers and the observations of the first author, as is highlighted in Table 2.
Whilst not directly comparable to the observations made by the first author (as teachers did not make observations per se, rather they rated on a scale from "never" to "all the time"  the degree to which they perceived NAP and non-NAP students played together), these findings demonstrate a significant gap between the perceptions that teachers have of instances of shared play and those observed by the first author.
For instance, a third of all teachers perceived that NAP and non-NAP students frequently play together in an integrated fashion on the oval and basketball, whereas the first author never witnessed this occuring.Whilst it is of course possible that such integrated play does happen, and that it simply did not occur when the first author was conducting observations, this would seem unlikely as observations occurred over an eight week period.A more likely answer is that teachers' perceptions of NAP and non-NAP students playing together were actually instances of both groups playing in the same area, but not together.

Hills Primary School
In comparison to Plains Primary School, far fewer instances of NAP and non-NAP students playing in the same area (either together or separately) were observed by the first author, as can be seen in Fig. 4. Yet regardless of there being far fewer instances of playing in a shared space at HPS (only 24% of all instances of play observed at HPS were in a shared space, compared to 54% at PPS), there were still minimal instances of this actually constituting interactive play (4%), as opposed to both groups simply sharing the area (20%).It must be acknowledged here that the dearth of observations of NAP and non-NAP students playing in the same area may well have been due in part to the fact that HPS has a much larger play area overall in comparison to PPS, and many more distinct play areas than PPS.This may have meant that there were more opportunities for NAP students to gather together in the same area away from non-NAP students.This can be seen in relation to areas such as the sandpit and the main oval (where only NAP students were observed playing), as outlined in Table 3.
A log-likelihood ratio test found that there were indeed significant differences between NAP and non-NAP use of play areas, LRR (9, N = 37) = 38.92,p < .01.
Field notes recorded by the first author noted interesting similarities between the two schools, despite the marked differences in the amount of play in the same area occurring between NAP and non-NAP students.For instance, some NAP students were witnessed playing at the edges of main play areas on repeated occasions, as this extract from the field notes indicates: "I ask the NAP girls what they are going to do today, and they say they are just going to 'hang around'.Later I see them walking around in the edge of the oval".Again, these appeared to be younger students, and more often than not these were girls.Whilst this was not observed as often as at PPS, it was still noticeable that despite considerably more space in which to play and main play areas to utilise, NAP students still played at the edges.Another similarity between the two schools was that younger NAP students at HPS were also often observed eating lunch and then playing outside of their classroom, rather than in other play areas.This tendency was rarely, if ever, seen to be the case with non-NAP students.An excerpt from the field notes illustrates this play pattern of NAP students at HPS well, and echoes those reported earlier from PPS: While I watch the NAP children whose class I was in today, they stay out the front of their class and finish off their lunch and don't run off anywhere.I have noticed this a bit -that NAP children have been hanging out the front of their rooms finishing off food -I haven't seen this with children from mainstream classes.
Likewise, NAP students at both HPS and PPS were observed playing across class groups with students in non-NAP classes who had previously been in NAP classes.This would appear to indicate that NAP students who have transitioned to non-NAP classes prefer to play with their old classmates rather than their new ones in their non-NAP class.Due to the short timeframe of this study we do not have data regarding how long this trend lasted (i.e., we are unable to say how long it generally takes a NAP student who has transitioned out of the NAP to start to feel 'at home' in the school space).Clearly, however, this transition is a gradual one.
Similarities were also found between the schools in terms of rules which potentially functioned to disadvantage NAP students, For example, an incident was witnessed at HPS in regard to two different-aged siblings from NAP classes who preferred to play together, but who were discouraged from doing so by yard teachers who felt that students should play in age-appropriate groupings.Again, this demonstrates some of the ways in which educators may police play spaces to the disbenefit of NAP students who may already feel marginalised within the school environment, and who may experience further marginalisation when segregated unnecessarily from one another.
An interesting difference was noted between the two schools, namely that there was one particular space -the sandpit -where NAP students at HPS were seen to play exclusively, as can be seen in Table 3.It may be suggested that this area could be claimed by NAP students primarily due to the larger number of available areas (thus resulting in less competition for space).Another reading of this could be that the sandpit could be claimed by NAP students as it was located directly outside a NAP classroom (see Fig. 5), and this reading is in line with previous observations regarding the tendency for NAP students to remain close to their class spaces.
In terms of thinking about the claiming of space as directly related to possessing the power to do so, this latter explanation may more adequately account for the predominance of NAP students inside the sandpit (and non-NAP students outside of it) in terms of proximity to their class and the security it offers, rather than the ability of NAP students to assert authority over an area per se (especially as it was typically younger students witnessed playing in the sandpit).As with teachers at PPS, teachers at HPS reported the perception that interactions between NAP and non-NAP students did occur, with the majority of teachers reporting that interactions occurred at least "sometimes" (54%), but that this also occurred in some areas "frequently" (21%), as Fig. 6 indicates.
As a whole, the perceptions reported by teachers would suggest a far greater degree of interaction than was observed by the first author.This can be seen most clearly if we are to compare specific perceptions of the frequency of interactions as reported by teachers for each of the main play areas in Table 4 with the observations made by the first author as reported in Table 3.
Again, whilst these perceived levels of interaction reported by teachers cannot be directly compared to those produced by the first author, the differences are startling.For example, almost a third of all teachers at HPS suggested that they had observed interactions occurring between NAP and non-NAP students "frequently" on the oval, in the sandpit, and on the play equipment spaces.By comparison, the first author made no observations of actual interactions occurring (rather than a shared use of space) between NAP and non-NAP students on the main oval or in the sandpit, and only one observation of interactions occurring on the upper primary play equipment.Similarly, almost one fifth of teachers perceived interactions occurring between NAP and non-NAP students "all the time" in the play equipment spaces, whereas the first author rarely observed this.
Again, we would suggest that the differences in these levels of perceived interaction may signify different meanings given to the use of yard spaces by teachers as compared to the first author.Whilst the first author paid close attention to actual interactions between the two groups (as opposed to both simply being in the same area), teachers may not have noticed this difference.Whilst it is fair to acknowledge that teachers do not have the same luxury of time as does a researcher who is not charged with the task of keeping students safe in the play area, it is nonetheless significant that teachers perceived more interaction than perhaps does occur.In other words, if teachers believe that interactions already occur, they may put less work into further encouraging positive interactions, thus allowing the potentially low current rates of interaction to continue.

Discussion
As the findings presented here suggest, little interaction was observed between students in NAP classes and students in non-NAP classes during the time in which the observations at the two primary schools occurred.This lack of interaction was evident even when NAP and non-NAP students were seen playing in the same space.As such, students currently in NAP classes appeared to be largely separated from the non-NAP students in the school, whose class and peer groups many would ultimately be entering.Similarly, the findings show that (with few exceptions) NAP students appeared to have difficulty claiming spaces within the schoolyard as their own, meaning that they tended to play on the peripheries of the playground spaces.In the remainder of this discussion we consider these findings and further explore some possible reasons for why such segregation occurs in the schoolyard, along with offering some suggestions for ways in which schools could set out to address the spatial divisions noted between NAP and non-NAP students.
Before doing so, however, we would like to acknowledge here once again that due to the fact that the research involved primary school students within a school environment, adult/child divisions in the distribution of power were present in the study (see Darbyshire, MacDougall and Schiller, 2005;Smith andBarker, 2000 andThomson, 2005 for a discussion of the unequal power relations between children and adults, and Van Ausdale and Feagin, 2001:40 for a discussion of the "least-adult" approach to research).As mentioned previously, informed consent as it is understood for adults was not obtained from the students at the schools we studied (though their parents were made aware of the research).Furthermore, and due to the injunction to remain a neutral observer (Van Ausdale and Feagin, 2001), the first author did not intervene in moments when she noted instances of marginalisation (though these were reported to teachers).As such, and as researchers committed to political change through our research findings, we acknowledge our own complicity with existing power structures that further marginalise NAP students within schools.Furthermore, we would like to acknowledge our position here as academics, rather than as primary school educators.In so doing we recognise the difficulties faced by teachers in relation to funding complexities and other issues which may undermine their capacity to work more concertedly in ensuring the inclusion of NAP students.As such, we do not wish to be overly critical of the ways both schools deal with NAP students, but rather offer these suggestions by way of building upon the good work being done already (see Van Ausdale and Feagin for a detailed discussion of the ways in which racial tensions may still exist in schools or preschools despite best-practice anti-bias curriculum and policy).
In regards to the implications of our findings, it is noteworthy that whilst this study did not set out to compare any differences that may have existed between the two schools, some important variations in use of space at the schools were observed.These are worth mentioning due to the fact that they highlight the argument that different school geographies will quite possibly lead to different patterns of play amongst students (Miller, 1997).As mentioned previously, our data showed that NAP students at HPS were able to claim some spaces within the school grounds for their almost exclusive use in a way that was not seen at PPS.The most obvious reason for this difference between the schools was simply that HPS had a much larger, more spread out playground space in which there was a larger number of easily definable sections for students to play in.These differences in the geographies of the schools are important since, as previous research has noted, power relations between students are often at play within playground and other school spaces (Van Ingen and Hallas, 2006).This can be seen particularly in divisions of power between students from mainstream groups and students from ethnic backgrounds which might be considered "other" (Kuriloff and Reichert, 2003).As such, we would recommend that, where possible, the need for larger playground spaces is taken into account when NAPs are established in new schools in order to ensure that NAP students have the possibility of claiming some spaces as their own.For current schools, we would suggest that it is important to consider how the future allocation of more NAP students to existing NAP schools may detrimentally impact upon the current use of limited spaces within NAP schools with smaller yards.
Despite differing in their perceptions of the use of space from those observed by the first author, teachers' opinions at both schools were remarkably similar.For example, in casual discussions held with teachers by the first author, as well as in data obtained from the questionnaires, many teachers at both schools suggested that the segregation of NAP and non-NAP students in the playground was due to the fact that students in both NAP and non-NAP classes tended to play within their class groups.However, as discussed in the analysis section, NAP students at both PPS and HPS were frequently seen playing across class groups rather than simply within their own classes.Echoing previous literature (Van Ingen and Hallas, 2006), cross-class friendships were especially noted in large groups of boys from NAP classes playing soccer or some other sport on the oval.At HPS these boys frequently occupied the oval on their own, with boys from non-NAP (almost all of whom had not previously been in a NAP class) utilising other, usually paved, areas for their sports games.At PPS, boys in NAP classes were also seen on the oval, however here it tended to be on the peripheries as a large group of boys from non-NAP classes were frequently seen in the centre of the oval here playing cricket around the pitch.
Whilst these examples refer solely to boys' use of space, girls from NAP classes were also frequently seen playing across classes, and many indicated in conversations with the first author that they were friends with students from other NAP classrooms.Such friendships were not necessarily restricted to similar cultural backgrounds, as was sometimes suggested by teachers.Whilst this was occasionally seen to be the case (especially amongst students who had only very recently arrived in Australia and did not speak any English), NAP students' friendships extended across NAP class groups to include students from a diverse range of cultural backgrounds.However, such cross-class interaction did not extend to interaction between NAP students and students from other non-NAP classes, except in instances in which, for example, a NAP student had a sibling in a non-NAP class with whom they frequently played.As such, it seemed that NAP students across all age levels and backgrounds "stuck together" rather than "integrated" with non-NAP students.
Interestingly, conversations with NAP teachers indicated that programmes were run at both schools during lunch times for NAP students to participate in.These programmes were designed to encourage inter-class interaction amongst NAP students and to provide students with the social skills that the teachers felt were necessary for students from refugee and other migrant backgrounds.However, similar programmes were not run with the aim of building bridges across NAP and non-NAP classes (although HPS indicated it was about to start a course designed to encourage NAP/non-NAP interaction).Several NAP teachers at both schools also indicated that they had difficulty organising non-NAP teachers to take a student who was about to transition into a non-NAP class into their room for lessons such as art or sport.This was the case even when the NAP teacher indicated that the student did not have to be graded and would not place any extra responsibility on the non-NAP teacher.As such, cross-class interactions between NAP and non-NAP were reported as being difficult to organise and rare, meaning that the separation of NAP and non-NAP students extended beyond just the playground.As such, it is important to consider how this lack of provision of opportunities for cross-class interactions on the part of teachers may be seen to model for students a separatist approach to NAP/non-NAP interactions in the playground.Interestingly, the possible impact of lack of in-class contact was acknowledged by the NAP Vice-Principal at PPS, who indicated that she had to lobby very hard in order to be able to take both NAP and non-NAP students from the upper primary school on camp together, as she felt that this shared experienced would increase the levels of interaction between NAP and non-NAP students.
Of course there are yet other reasons as to why the separation between NAP and non-NAP students in the playground potentially occurs.The first of these is that instances of racism or bullying in the school yard could lead to a lack of interaction between NAP and non-NAP students.In questionnaire responses, 75% of teachers across both schools reported witnessing instances of racism in the schoolyard, and of these instances 50% were reported as being enacted against NAP students.If these figures are indeed representative, NAP students are likely to feel highly marginalised within school spaces.One implication of this is that NAP students may perceive a requirement to play separately to non-NAP students who are in a position to "set the rules" for playground interaction or perpetuate marginalisation.This was seen in some instances observed by the first author in her ethnographic observations, including instances of NAP students saying that they had to leave an area as other (non-NAP) students had come along, or that they wanted to use the play equipment space but only could if they were able to "get there first" and lay claim to the space as theirs.
Given the importance of ensuring that refugee and other NAP students are not marginalised within school grounds, we would suggest that it is a responsibility of schools to ensure that positive interactions between NAP and non-NAP students do take place, and that NAP students are not isolated or marginalised within the school grounds (Morrice, 2007).One way in which interactions between NAP and non-NAP students could be achieved is through the use of shared class space (Miller, 1997).Ensuring that a number of classes which do not rely on English skills are shared by students from NAP and non-NAP classes could encourage students from both groups to interact together.For example, several NAP teachers commented that they felt that their students could express themselves well through art, and that they really loved time spent painting or drawing.Having combined art classes could provide a point of contact between NAP and non-NAP students in which the skills of NAP students are highlighted.Combined classes also provide an opportu-nity for the school community to not simply "integrate" NAP students in a top down approach which prioritses "norms" of the playground space, and allows refugees and other NAP students an opportunity to demonstrate their own wealth of knowledges, experiences and skills (Morrice, 2007).
Another suggestion that we offer is for all teachers to be aware of differences which may exist for NAP students (especially newly arrived NAP students, and those with experiences of forced migration) and for schools to be prepared to adjust playground rules accordingly.As such, we suggest that schools ought to be prepared to change school rules which contribute to making spaces centred around dominant values or norms, and which further disadvantage NAP students in the school space.For example, rules regarding students of different ages being unable to play together (which were in place in both the schools in this study) prevent siblings playing together for support in the school yard.Recognising rules which disadvantage NAP students in this way and changing them for all students in the school space would increase the likelihood that NAP students feel "at home" in the school yard, as well as centring some of the family and community practices of refugee and migrant children.In other words, understanding the cultural norms through which newly arrived students have previously been operating (i.e., multi-age sibling play may be a standard cultural practice for some students), as well as any behaviour traits which they have acquired to cope with situations very different to a South Australian primary school environment (such as playing together for safety reasons in countries of origin in a state of war), could help schools to create playground spaces more suited to the needs of NAP students (Peirce, 1995).
Again, this is an understanding which is required by all teachers, not just those teaching NAP classes.Questionnaire responses by NAP teachers indicated that they frequently felt that non-NAP staff members did not understand the complexities faced by NAP students, and that they needed to spend more time "getting to know" NAP students transitioning to their classes.The lack of understanding of this requirement was also reflected in some non-NAP teacher responses, which asked for less information to be provided about NAP students as the documents they were currently given were long and involved.Clearly time constraints impact on the ability of teachers to invest time into understanding NAP students' backgrounds, however we would suggest that such a time commitment is necessary.Therefore, we propose that schools invest in providing opportunities for staff members to liaise across NAP and non-NAP classes in order to better understand specific requirements which NAP students face (especially those with refugee backgrounds).
Similarly, ensuring that schools norms are not centred around those of its "mainstream" families will help make the school environment a place in which NAP students feel welcomed.For example, both schools celebrated Easter, even though significant numbers of its students are not from Christian backgrounds.Although such celebrations centred around chocolate and bunnies rather than Christian iconography, they nonetheless reinforced norms and values which were different from those brought by most NAP students.This was acknowledged by class teachers who, in one instance, had not cut out paintings designed to be Easter eggs into egg shapes, and had not sent them home with students before the Easter break as "its not something they do at home anyway -just at school".We would argue that such practices are not just "a bit of fun", and in fact reinforce to many students that they are entering an environment which is not their own, and in which they must conform to values held by other students.This issue of certain dominant cultural values being privileged above others was also evident in the allocation of space for Muslim students within the schools, or more precisely, the lack of allocation of space for a prayer room at one of the schools.Whilst the school said they provided areas for Muslim students to pray in if required, the lack of an appropriate designated space set aside for these students sends a message to families that such practices are not the norm, and are not valued enough by the school to provide a permanent space.
As such, a focus on recognising the values, customs, skills and knowledges brought by NAP students to the school, together with ensuring that the school environment supports the needs of students from all backgrounds, would also mean that NAP students may be more likely to see the school space as their own, rather than a space into which they must attempt to "fit into".The rhetoric of "inclusion" and "integration" frequently referred to by teachers effectively places the onus on NAP students to conform to school customs and expectations, rather than creating an environment in which NAP students are able to shape the space to the same extent as non-NAP students.Acknowledging power relations between NAP and non-NAP students, creating spaces within structured school time to encourage interaction between NAP and non-NAP students, and providing spaces which focus on the differing knowledges of NAP students (and which will therefore also provide for the ongoing needs of students exiting the NAP into non-NAP classes) will all help to create an environment in which primary schools are not simply seen as solely the space of mainstream students into which NAP students must somehow find a way to "fit in", but rather as spaces shaped through the coming together of a range of differing groups (Morrice, 2007).
In saying this, our suggestion here is of course not for a version of "melting pot" multiculturalism where power differentials are ignored.Schools must acknowledge that despite their best efforts, dominant cultures will likely continue to dominate, to the disbenefit of marginalised cultures.As such, we recommend that schools are active in helping students to explore the cultural and social geographies of their schools in order to become more empowered to take a critical stance to their own education (Kuriloff and Reichert, 2003).However, it is likely to be the case that many non-NAP students and their families will continue to view NAP students and their families as strange or unwelcomed others who are expected to prove their worth in order to be accepted.Given that school spaces are not discrete, but bounded to the community in which they are placed (Holloway and Valentine, 2000), school environments have been identified as contributing to this logic (Lovell and Riggs, 2009), and resisting this requires active attempts by schools to challenge a logic of inclusion that is 1) set on the terms of the dominant group, and 2) requires marginalised groups to prove their worth.Instead, our suggestions here have been to encourage both recognition of the power differentials that exist (and which are likely to continue), but to nonetheless attempt to afford opportunities for NAP students to determine their own terms for engagement that recognise their complex needs and histories.

Figure 1 .Figure 1 .
Figure 1.Distribution of total observations of student play by group at PPS

Figure 3 .Figure 3 .
Figure 3. Teachers' perceptions of joint play occurrences across all locations at PPS

Figure 4 .
Figure 4. Distribution of total observations of student play by group at HPS.

Figure 5 .
Figure 5. Map of Hills Primary School.

Figure 6 .Figure 6 .
Figure 6.Teachers' perceptions of joint play occurrences across all locations at HPS

Table 1 .
Distribution of observations of student play by location at PPS.

Table 2 .
Teachers' perceptions of joint play between NAP and non-NAP students by location at PPS.
igure 4. Distribution of total observations of student play by group at HPS F

Table 3 .
Distribution of observations of student play by location at HPS.

Table 4 .
Teachers' perceptions of joint NAP and non-NAP play by location at HPS.